The neural dynamics of integrating prior and kinematic information during action anticipation in sport

• EEG and MVPA distinguish how athletes and controls integrate information. • Experts rely more on kinematic information; controls lean on prior information. • Temporal dynamics of neural decoding highlight strategy differences between groups. • Experts showed enhanced alpha ERD during kinematic processing. • Sensitivity to prior–kinematic mismatches increased as the task progressed. Effective action anticipation in sports hinges on the integration of prior knowledge and kinematic cues, enabling athletes to respond swiftly and accurately in real-time scenarios. However, the neural mechanisms supporting this integrative process remain insufficiently understood. This study addressed this gap by using electroencephalography (EEG), combined with both multivariate and univariate analyses, to investigate how expert basketball players and non-athlete controls process prior and kinematic information during a sport-specific action anticipation task. Eighty-five participants (44 experts and 41 controls) were asked to predict the outcomes of basketball free throws presented via video clips, either with or without outcome-based prior information cues. Multivariate pattern classification and contingent negative variation (CNV) analyses revealed distinct anticipatory strategies between groups, with experts predominantly relying on kinematic information, whereas controls showed greater sensitivity to prior information. Additionally, time–frequency analysis of alpha-band activity indicated stronger desynchronization in experts, reflecting enhanced cortical engagement during kinematic processing. Notably, alpha ERD was significantly stronger for incongruent trials in the later phase of the task, suggesting increased cortical engagement when resolving conflicts between prior expectations and observed actions. These findings advance our understanding of the temporal dynamics and neural mechanisms underlying action anticipation, and highlight the value of combining EEG with multivariate decoding approaches to characterize individual differences in predictive processing.